Distinct electroencephalography alpha-band rhythms map macroscale brain gradients

Abstract

Abstract Brain gradients provide a low-dimensional framework for describing cortical organization, linking functional connectivity, anatomy, development, and transcriptional architecture. Whether this macroscale organization also structures fast electrophysiological dynamics remains less clear. Here we used the longstanding observation that alpha-band activity is not unitary, and that some individuals exhibit multiple alpha peaks, as an entry point for examining cortical gradients in resting-state electroencephalography (EEG). We analyzed source-reconstructed EEG from 608 adults, comprising 3,255 recordings acquired across eyes-open and eyes-closed rest before and after cognitive testing. Within individuals, parcel-wise alpha peak frequencies frequently departed from unimodality, revealing 2 separable alpha-frequency components in 38.3% of recordings. These components were not spatially arbitrary: frequency-defined clusters aligned with canonical functional networks, most prominently separating visual and somatomotor systems, although their relative frequency ordering varied across individuals. Low-dimensional analysis of cluster network composition further identified 2 principal axes: a sensory-differentiation axis contrasting visual and somatomotor contributions, and a sensorimotor–association axis resembling established functional-connectomic gradients. These findings suggest that split alpha reflects multiple sensory-anchored oscillatory modes embedded within macroscale cortical gradients, linking a classical EEG phenomenon to the brain’s hierarchical architecture.